ARM-software/ComputeLibrary
The Compute Library is a set of computer vision and machine learning functions optimised for both Arm CPUs and GPUs using SIMD technologies.
C++MIT
Issues
- 5
- 5
Can Khadas Edge2 board be supported?
#1056 opened - 2
Differences between run() and run_op()
#1055 opened - 5
- 14
How to build universal ARM Compute?
#1053 opened - 1
Is Weights Manager supported?
#1052 opened - 3
- 2
- 5
- 1
- 1
Wrong values for NEElementwisePower kernel
#1047 opened - 2
- 3
Low performance with NETranspose on aarch64
#1045 opened - 5
NEReduceMean accuracy issue for NHWC
#1044 opened - 3
- 1
- 6
- 1
- 6
- 7
- 4
Problem with SpaceToDepth layer
#1036 opened - 0
Got a problem about the batchnorm's codes
#1035 opened - 6
GPU Convolution for INT8 is slower than FP16
#1034 opened - 1
- 4
Postop examples with NEON target
#1032 opened - 7
- 3
- 1
NEScale: Cubic interpolation
#1029 opened - 4
Build ERROR (VFP register)
#1028 opened - 3
NEGEMMLowpMatrixMultiplyCore support type
#1027 opened - 3
- 2
- 4
- 14
- 1
Usage of optimized-routines functions
#1020 opened - 2
A question in Deconvolution
#1019 opened - 1
clfunctions keep using cpu device
#1018 opened - 1
Result of Deconvolution not correct
#1017 opened - 0
- 1
Example to use the bfloat16 kernel directly.
#1015 opened - 6
Extend RNNLayer
#1014 opened - 8
- 4
- 1
- 3
- 7
- 2
- 1
- 2
Int8 GEMM kernel
#1006 opened - 4
build errror for RK 3588S Ubuntu Jammy
#1005 opened